App Builder is the AI-agent-powered custom app surface inside Vortex Apps. You describe what you want your workflow to do in plain English. The agent parses the description, identifies the data sources and actions required, generates a structured plan for your review, deploys the app to a sandbox for testing, and promotes it to production once you approve. The path from idea to running app is measured in minutes, not weeks of developer time. App Builder replaces the legacy Agent Hub that shipped in Vortex IQ V1. The Agent Hub model was a curated catalogue of pre-built per-platform recipe agents — useful, but rigid. App Builder is the V2 evolution: instead of picking from a catalogue, you describe what you need and the agent assembles a bespoke app. The legacy recipes still live inside App Builder as starter templates, but they are examples of what is possible, not the only choices on offer.Documentation Index
Fetch the complete documentation index at: https://docs.vortexiq.ai/llms.txt
Use this file to discover all available pages before exploring further.
What App Builder is for
App Builder is for workflows specific to your business that no off-the-shelf app covers. Common patterns:- “Every Tuesday at 9am, summarise last week’s sales by SKU and post the top 10 to the #ops Slack channel.”
- “When a customer places their third order, send them a personalised thank-you email with a discount code.”
- “When checkout conversion drops more than 10% week over week, open a Jira ticket and assign it to the engineering lead.”
- “Every morning, list the SKUs that are out of stock on Amazon but in stock on Shopify, and email the inventory manager.”
- “If a Datadog alert fires on the storefront, post the incident to Microsoft Teams with the latest Vortex Mind report attached.”
How it replaces Agent Hub
The V1 Agent Hub shipped pre-built agents for specific platform tasks. The catalogue included agents for Shopify (AI Growth Advisor, Discount Creator, Image Agent, KPI Agent, Pages SEO Agent, Product SEO Agent), BigCommerce (Image Compression, Image Optimisation, KPI Agent, Product SEO Agent, StagingPro Migration Agent), Adobe Commerce (Image Optimisation Agent, Pages SEO Agent, Product SEO Agent), and cross-platform workflows (Ask Viq Agent, Brand DNA Agent, Campaign Manager, Google Ads Performance, Competitor Agent, and others). Each was a fixed agent — you picked the recipe, you ran it, and you got the result the recipe was built for. Customising required engineering. App Builder flips the model:| V1 Agent Hub | V2 App Builder | |
|---|---|---|
| How you find what you need | Browse a catalogue of pre-built agents | Describe what you want; agent assembles it |
| Customisation | Limited; engineering required | Plain-English edits at any stage |
| Tool availability | Per-agent fixed toolkit | Full toolkit available; agent picks the right tools |
| Platform scope | Each recipe was platform-bound | Apps span multiple platforms natively |
| Schedule | Manual run in most cases | Scheduled, event-triggered, or on-demand |
| Sandbox | Not first-class | First-class staging surface before production |
| Deployment | Run from catalogue UI | Owned, editable production app |
Building an app
Open App Builder and start a new app
Navigate to Vortex Apps → App Builder in the Vortex IQ console, then click + New App. The prompt panel opens with placeholder text guiding you: describe what you want the app to do, mention triggers, data sources, and where the output should go.
Type your prompt
Write a description of your workflow in plain English. A good prompt has four parts:
- Trigger — “Every Tuesday at 9am” / “When checkout conversion drops 10%” / “When a customer’s third order ships”
- Data — “Shopify sales” / “Google Ads campaigns” / “Amazon inventory in CloudHub”
- Logic — “Group by SKU” / “Sort by revenue” / “Filter for the last 30 days”
- Output — “Post to #ops in Slack” / “Email the inventory manager” / “Open a Jira ticket”
Review the agent's plan
After a few seconds the plan view opens. The agent generates a structured plan covering six parts: trigger, data sources, logic, output, schedule, and approval gates. Each row is editable inline. Refinements are conversational — tell the agent what to change (“use the last 30 days instead of last 7”, “add a conversion rate column”, “send to a different Slack channel”) and the agent regenerates the plan after each edit.
Deploy to sandbox
Click Deploy to sandbox. The app provisions a sandbox runtime in 10–30 seconds. From the sandbox view, click Run test now. The agent connects to your data sources, applies the logic, and routes write actions to safe channels — a sandbox Slack channel, your inbox, a sandbox Jira project — so production systems are not touched. The actual output renders inline.
Verify the test output
Review the output. If something is wrong — the agent included refunded orders when you wanted completed only, or the column order is off — go back to the plan, add the clarifying instruction, and re-deploy to sandbox. Iteration is fast because each round takes seconds. You can also switch to the Inspect tab to view the agent’s full reasoning trace, the tool calls it made, and the data it read.
Promote to production
When the sandbox output is correct, click Promote to production. Write actions now route to real channels. The next scheduled run shows in the header. Every production run is logged to Vortex Memory with the input data, the agent’s reasoning trace, the tool calls it made, and the output it produced. You can pause the app, edit it, view error history, and roll back to a previous version from the Deployment and Monitoring view.